Large-area diamond thin film on Q-carbon coated crystalline sapphire by HFCVD

Large-area diamond thin film on Q-carbon coated crystalline sapphire by HFCVD

Journal of Crystal Growth 504 (2018) 17–25 Contents lists available at ScienceDirect Journal of Crystal Growth journal homepage: www.elsevier.com/lo...

2MB Sizes 0 Downloads 19 Views

Journal of Crystal Growth 504 (2018) 17–25

Contents lists available at ScienceDirect

Journal of Crystal Growth journal homepage: www.elsevier.com/locate/jcrysgro

Large-area diamond thin film on Q-carbon coated crystalline sapphire by HFCVD

T



Ariful Haque , Punam Pant, Jagdish Narayan Materials Science and Engineering, North Carolina State University, Raleigh, NC 27695.7916, USA

A R T I C LE I N FO

A B S T R A C T

Communicated by T. Paskova

The growth of diamond on transparent substrates like sapphire presents a great challenge because of the large thermal misfit between the film and the substrate, absence of any carbide layer during diamond growth, and low nucleation density during chemical vapor deposition (CVD) growth process. In this study, we report on the use and the role of Q-carbon as an intermediate layer to successfully deposit large-area diamond film on c-sapphire by hot filament chemical vapor deposition (HFCVD). The Q-carbon consists of very high-density diamond tetrahedra which act as the embryo for diamond nucleation. Different techniques such as X-ray diffraction, scanning electron microscopy, and Raman spectroscopy show that continuous diamond films with good crystallinity and without any impurity phase can be deposited on the Q-carbon coated single crystal sapphire substrate. The Q-carbon layer is very adherent and it negates the thermal mismatch between the diamond film and the sapphire substrate. A small blue shift in the Raman peak of the diamond from its equilibrium position suggests the deposition of the CVD diamond film with minimal stress (1.14 GPa). This technique of growing large-area continuous diamond thin film with excellent crystalline quality on a single crystal sapphire substrate can serve as a platform for the development of next-generation corrosion and erosion resistant infrared windows, state-of-theart optoelectronic devices, and advanced scanning probe microscopy systems.

Keywords: A1 Stresses B1 Diamond B1 Sapphire A3 Chemical vapor deposition

1. Introduction Diamond thin films have unique physical and chemical properties which are needed for different technological applications of immense commercial importance. Especially, the deposition of large-area diamond film on single crystal sapphire holds a lot of promise in various technological applications due to unique properties of the substrate, which include good optical transparency, high-melting point, high hardness and chemical inertness, etc. The applications of diamond on sapphire include diamond polishing wheels, corrosion and radiation resistant infrared windows, solid-state lasers, novel scanning probe microscopy applications such as nanoindentation and high-resolution imaging of soft samples, and so on [1,2]. In fact, a model laser gain element formed by integrating diamond and a solid-state laser material, such as Ti: Sapphire, has been designed to obtain an increased average output power up to a megawatt level [3]. All of these technological milestones can be achieved by establishing a successful deposition route to grow large-area and high-quality diamond on single crystal sapphire. Although many comprehensive studies about chemical vapor deposition (CVD) of diamond on different substrates have been conducted over the past few decades, relatively little effort has been put on the



development of large-area CVD diamond on single crystal sapphire substrate due to different technical obstacles. The utilization of sapphire as a substrate for CVD diamond growth is difficult due to low thermal conductivity (27.2 W/m-K) of sapphire, huge difference in thermal expansion coefficients between the diamond film and the sapphire substrate, low nucleation density, and poor adhesion [4,5]. Since the Gibbs free energy of formation of aluminium carbide (−32.1 KJ/mole) is much lower compared to that of aluminium oxide (−1675 kJ/mole), the carbide layer for the diamond nucleation is not favored [6]. Over the years, attempts to grow large-area diamond thin film on transparent and hard substrates have achieved, at best, limited success. Thermodynamically, sapphire is a very stable material, and its conversion from Al2O3 to an intermediate carbide layer during the fabrication of CVD diamond is unfavorable [7]. The absence of any carbide layer at the interface makes the diamond film poorly adherent on the sapphire substrate. The adherence is even worse due to the large thermal expansion coefficient mismatch between the sapphire and the diamond, which leads to a huge thermal stress (compressive) in the diamond film during the cooling process. To overcome adhesion and stress-related problems, researchers have proposed different techniques, such as an in situ two-step hot-filament chemical vapor deposition

Corresponding author. E-mail address: [email protected] (A. Haque).

https://doi.org/10.1016/j.jcrysgro.2018.09.036 Received 29 June 2018; Received in revised form 7 September 2018; Accepted 23 September 2018 Available online 25 September 2018 0022-0248/ © 2018 Published by Elsevier B.V.

Journal of Crystal Growth 504 (2018) 17–25

A. Haque et al.

on sapphire without any cracking or delamination. The analyses from scanning electron microscopy (SEM), X-ray diffraction (XRD), and Raman spectroscopy show that the Q-carbon assists the diamond nucleation, and a large-area continuous diamond film with low compressive stress can be obtained. We have conducted a detailed analysis of stresses in these films and found that the Q-carbon seed interlayer helps to reduce the effect of the thermal misfit strains which arise as a result of the difference in coefficients of thermal expansion between the diamond film and the smooth sapphire substrate.

(HFCVD) [8], treating the substrate surface by in situ pre-deposition of a carbon layer to form aluminum carbide at the surface during the HFCVD, microwave plasma chemical vapor deposition (MPCVD) [9], and using the plasma flame CVD which supplies activated diamond species to the substrate in the plasma flame [10]. However, all of these studies yielded relatively low nucleation density (106 –107 sites/cm2), and formation of large-area continuous diamond thin films was not obtained. The dc-biasing technique has been successfully used to deposit diamond on mirror-polished Si substrate with improved nucleation. However, this method cannot be applied to the insulating sapphire substrate [11]. The growth of polycrystalline diamond films on single crystal sapphire using HFCVD has been studied [12]. Problems related to the stress generation, poor adhesion at the film-substrate interface, delamination, and film cracking were observed for deposited areas greater than ∼100 µm2. An alternative technique, by using patterned silicon on sapphire for CVD of diamond, was used to grow diamond on sapphire [7]. In this case, Si acts as a partial stress-relieving layer. However, the area of continuous diamond film greater than ∼200 µm2 still delaminated. These problems restrict the deposition of large-area diamond film on sapphire and put an upper limit even after using a crystalline Si layer as an intermediate layer. Diamond nucleation on mirror polished optically transparent substrates is another major challenge. To increase the nucleation density researchers tried faceting or mechanical scratch pretreatment, by abrading using ultra-sonication or annealing in air at high-temperature, on sapphire during the HFCVD of diamond thin film [13,14]. But these pretreatment techniques damage the surface of the polished substrate, and hinder practical applications of diamond on sapphire. In another study, researchers claimed that a high nucleation density and a rapid lateral growth rate enabled the diamond nuclei to coalesce and cover the entire surface of the substrate in a short time using HFCVD (after 20 min manual aberration of 1 μm diamond seeds on the sapphire substrate) [12]. However, further diamond deposition (4 h) caused a large part of the diamond film to delaminate. Recently, large area diamond thin films on sapphire have been synthesized using the adamantime coating on the substrate [15]. A good quality diamond film on sapphire was obtained only after a very long deposition time of 48 h. In this study, we have investigated a novel route to grow large-area CVD diamond thin film on sapphire by using Q-carbon as an intermediate layer to overcome the inherent problems related to the nucleation and growth of diamond on single crystal sapphire. We have successfully demonstrated a simple technique for the direct deposition of large-area, and excellent quality diamond thin films on the Q-carbon coated sapphire substrate by HFCVD. The Q-carbon is a new phase of carbon discovered by our group recently, which consists of diamond tetrahedra with a very high packing efficiency due to the ultrafast quenching from the liquid phase of carbon by nanosecond pulsed laser annealing [16,17]. This layer helps to grow a very large-area diamond

2. Experimental We have grown the large-area CVD diamond film using a two-step method, i.e. first, the formation of Q-carbon using nanosecond pulsed laser annealing (PLA) of the amorphous diamond-like carbon (DLC) film, and second the growth of diamond film on Q-carbon by HFCVD. To fabricate the Q-carbon, at first, we deposited ∼500 nm thick amorphous carbon thin film on the c-plane sapphire substrate using pulsed laser deposition. The substrate was cleaned with acetone vapor for 10 min, with methanol for 5 min by ultra-sonication, and dried with nitrogen gas. The amorphous carbon film was deposited in a stainless steel chamber evacuated to a base pressure of ∼1 × 10−6 torr. We used a pulsed krypton fluoride (KrF) excimer laser (λ = 248 nm, repetition rate 5 Hz/s, pulse width 25 ns, laser fluence of ∼2 J/cm2) during deposition. The as-deposited films were irradiated by an ArF laser pulse (wavelength = 193 nm, pulse duration = 20 ns) with an energy density in-between 0.6 and 0.7 Jcm−2 for conversion into Q-carbon. During the laser irradiation process, the as-deposited DLC film was melted in a super undercooled state and subsequently quenched to complete the whole process within 200–250 ns. After shrinkage from the PLD grown amorphous carbon layer, due to the super undercooling and subsequent quenching from the liquid state of carbon which is produced due to nanosecond pulsed laser irradiation, the generated Q-carbon layer thickness varies from 35 to 150 nm [18]. During HFCVD of diamond thin film, the Q-carbon coated c-sapphire was inserted into the HFCVD reactor. We used a set of 4 straight tungsten wires as hot filaments. The distance between the filament and the substrate was maintained at ∼6–8 mm. Diamond thin films were deposited for 6 h. The schematic diagram of the HFCVD reactor used to deposit diamond on Q-carbon/sapphire by employing methane and hydrogen gas mixtures is shown in Fig. 1(a). The tungsten filament was gradually heated to a high-temperature (∼2000 °C). The CVD growth was done under a flow rate of 2 sccm for methane and 100 sccm for hydrogen at a chamber pressure of 20 Torr. Pyrolysis of the gas mixture takes place inside the CVD chamber, which results in the generation of diamond-forming active gas species. Upon reaching the Q-carbon/sapphire substrate surface, these diamond-forming active gas species initiate the nucleation and the formation of diamond. The substrate was

Fig. 1. (a) A schematic cross-section of the HFCVD system used for diamond growth, (b) schematic of the physical and chemical mechanisms occurring in the HFCVD reactor during diamond thin film deposition. 18

Journal of Crystal Growth 504 (2018) 17–25

A. Haque et al.

connect with the diamond-forming gas species due to the presence of numerous diamond tetrahedra, and (ii) the regions of adsorption sites larger than the nuclei size for the subsequent crystal growth. Although considerable amount of work has been done in the past on the diamond nucleation mechanism during the HFCVD process, the detailed steps which can explain the nucleation of the first diamond phase are still unclear. The hot filament generates atomic hydrogen, which react with carbon containing species to reduce and connect it with diamond/diamond nuclei. The growth of diamond is affected by the generation of different hydrocarbon species in the gas phase at high-temperature inside the CVD chamber and adsorption of those species on the substrate surface via interfacial interactions. The Gibbs free energies at different stages depend upon the C–H bond energies in different CHX species, i.e. CH4, CH3, CH2, and CH, and formation energy of H2. The reactions initiate from the dissociation of the molecular hydrogen by the hot filament at a very high-temperature to form highly active hydrogen radical followed by several steps (by which the atomic carbon is generated via the formation of CHX species upon dissociation of methane) given below [19].

placed on a substrate holder, and the temperature of around 800 °C was maintained during growth. The set of multiple filaments and the rotation of the substrate stage during the deposition assisted the formation of a uniform diamond thin film over the large substrate area. Fig. 1(b) represents the schematic of the physical and chemical processes occurring inside the chamber and shows the dissociation of precursor gases at high-temperature. The absorbed reactive species coalesce and subsequently form diamond clusters on the substrate under favorable conditions, which are also shown in this figure. Raman spectroscopy and XRD are the two most powerful techniques to characterize diamond thin films in terms of structure and bonding, respectively. These techniques provide a fingerprint of the presence of diamond and amorphous carbon phases, and the stress characteristics present in the film. Raman spectra were obtained using a WITec confocal Raman microscope system (alpha300M) micro-Raman spectrometer (triple monochromator, gratings 2400 lines/mm) with an x100 objective which focuses the laser beam in a spot of 2 μm. All of the spectra were recorded by CCD matrix with a very high spectral resolution (< 300 nm). The excitation of the Raman scattering was performed by a 532 nm diode-pumped solid-state laser. To avoid any local heating of the sample, a low laser power of ∼40 mW was used. The Xray diffraction (XRD) was performed on the HFCVD diamond film using the Rigaku SmartLab x-ray diffractometer with a CuKα x-ray source (25KV, 15 mA). Diffraction data were collected with a 0.01 step size in the 2θ range of 30–100° and a constant dwell time of 0.5 s per scan point. Bragg-Brentano reflection geometry was used during the data acquisition process. High-resolution scanning electron microscopy (SEM) with the sub-nanometer resolution was carried out on the laser irradiated samples using the FEI Verios 460L SEM.

H2 → 2H

ΔG0 = −75 kJ/mol

CH4 + H → CH3 + H2

ΔG0 = −42 kJ/mol

CH3 + H → CH2 + H2

ΔG0 = 17 kJ/mol

CH2 + H → CH + H2

ΔG0 = −38 kJ/mol

CH + H → C + H2

ΔG0 = −84 kJ/mol

C + Cd → 2Cd

ΔG0 = −480 kJ/mol

At the end, generated carbon atoms are added to the diamond structure via the final reaction. Although the formation of atomic hydrogen and the formation of CH2 species are not thermodynamically favorable owing to the small positive changes in Gibbs free energy, the high-temperature environment ensures a thermodynamically favorable condition [19]. sp1, sp2 and sp3 bonded carbon species coexist on the substrate surface and the etching of the sp1 and sp2 bonded carbon are ten times faster than etching of the stable sp3 bonded carbon, which assists the growth of sp3 phase [20]. Around 104 activated hydrogen atoms are required to convert a carbon atom to the final sp3 bonded state, therefore, we have used a high hydrogen to methane ratio (50:1) [21]. At the final stage, the disordered domains of the sp3 bonded carbon convert into an ordered diamond structure with the same bonding type. The high density of the diamond tetrahedra in Q-carbon on the substrate surface helps to achieve a very high nucleation density during the diamond growth. The overall morphology of the obtained diamond thin films after 6 h growth is shown by the high resolution SEM images in Fig. 3(a)–(c) at different magnifications. No visible crack/delamination was observed in the film implying that the Qcarbon layer helps to overcome the huge thermal mismatch between the film and the substrate. Fig. 3(c) is the low magnification image of

3. Results and discussion 3.1. Nucleation, growth, and microstructure of diamond Nucleation of diamond requires the formation or preexistence of diamond nuclei which are the smallest stable diamond units that form on the substrate and then act as suitable sites for the subsequent diamond thin film growth. Understanding the diamond nucleation is crucial because it not only affects the final grain size and orientation of the diamond thin film but also the film thickness, morphology, uniformity, defect concentration, adhesion, surface roughness, and so on. Fig. 2(a) shows an SEM micrograph of Q-carbon on c-sapphire, which has been used for the HFCVD diamond deposition. In conventional diamond depositions, seeds of nano/micro diamond are dispersed on the substrate, which act as diamond nucleation sites during the CVD process. Using Q-carbon, we obtained a relatively high diamond nucleation density (around 109 cm−2) for diamond growth. Fig. 2(b) shows the SEM micrograph of diamond on the Q-carbon/c-plane sapphire surface after 3 h of HFCVD. The Q-carbon provides a high number density of nucleation sites by providing: (i) very active surface adsorption sites to

Fig. 2. SEM image (a) of Q-carbon on c-sapphire and (b) diamond growth on Q-carbon coated c-sapphire after 3 hours of HFCVD. 19

Journal of Crystal Growth 504 (2018) 17–25

A. Haque et al.

Fig. 3. (a)–(c) Large-area HFCVD diamond thin film (thickness ∼ 3 μm) on Q-carbon coated c-sapphire at different magnifications, and (d) HFCVD diamond on untreated c-sapphire under the same deposition conditions without Q-carbon. All depositions were done for 6 h.

The coefficient of thermal expansion of the diamond film (αd) is less than that of the substrates (αs) considered in Fig. 4(a). As a result, the film will be under compressive stress and the substrate will be under tensile stress after cooling. The thermal misfit strain (εT) in the diamond thin film plane on a particular substrate can be expressed by Eq. (1).

the HFCVD diamond thin film, which corroborates that large-area continuous diamond thin film has been grown using this novel technique. For comparison, we have also grown the diamond on a c-sapphire substrate without a Q-carbon layer under the same deposition conditions. The morphology of the diamond growth on sapphire without Q-carbon using the same HFCVD conditions for 6 h is shown in Fig. 3(d). This image shows low diamond nucleation density, where the area of the substrate surface is barely covered with diamond crystals. Thus, we can conclude that the Q-carbon layer can enhance the coverage during diamond deposition. This is due to the presence of the diamond tetrahedra in the Q-carbon with a high packing efficiency, which can act as embryos for the diamond nucleation. The residual amorphous carbon is etched off by the reactive species. This etching of the amorphous carbon (driven by the highly reactive atomic hydrogen) occurs via the following thermodynamically favorable reaction [19]. Ca + 4H → CH4

εT = (αd−αs )ΔT

(1)

where ΔT is the change in temperature and is negative during cooling down. Based on the thermal properties, the thermal strain of the diamond film on sapphire, AlN, and silicon substrates at different growth temperatures are illustrated in Fig. 4(b). The thermal strain that can be generated in the diamond film plane on Si or AlN is much less than that on the sapphire substrate over the entire temperature range. Therefore, the high-temperature growth of diamond on Si or AlN is relatively easier from the thermal strain viewpoint. This plot shows that at 800 °C deposition temperature, a compressive strain of ∼0.43% is expected to be generated in a diamond thin film on the c-sapphire substrate. Such a high compressive strain due to the thermal misfit can generate ∼5.1 GPa stress in the film. The diamond film on sapphire can easily delaminate under such large amount of strain. Although researchers have been able to grow diamond on c-sapphire by electron–cyclotron–resonance chemical vapor deposition process using dispersed–particulate diamond suspension for nucleation at a lower temperature (500–550 °C), however, such low growth temperature is enough to generate approximately 3.5 GPa stress [24]. This amount of compressive stresses may cause delamination in the weakly bonded diamond film from the sapphire substrate. Although the optical evaluation of the continuous diamond film did not show any visible cracking or delamination, however, the edges of the film peeled off due to the built-in stress generated by the thermal misfit.

ΔG0 = −470 kJ/mol

3.2. Effect of thermal stress Fig. 4(a) represents the percent of the linear thermal expansion of sapphire, AlN, silicon, and diamond in the temperature range between 0 °C and 1200 °C, which helps us to understand the effect of thermal stress on diamond film at different growth temperatures [22,23]. The relative thermal expansion at different temperatures also implies the magnitude and nature of the thermal stress in the diamond film grown at a particular temperature. To plot this figure, we have determined the percent of the thermal expansion by integrating the corresponding expression for the thermal expansion coefficient of each of the material over the temperature range. From Fig. 4(a) we can observe that the percent of thermal mismatch is the highest for the diamond-sapphire system compared to other systems at the deposition temperature (800–850 °C). Therefore, the growth of diamond on sapphire at such high-temperature has always been a challenging task. The thermal misfit strain occurs due to the differences in coefficients of thermal expansion between the diamond thin film and the substrate.

3.3. Raman spectroscopy Raman spectroscopy is a very effective technique to characterize the diamond film for obtaining bonding characteristics. Fig. 5 shows the Raman spectra of the diamond film grown on the c-sapphire substrate. This spectrum shows a sharp distinct characteristic Raman peak of 20

Journal of Crystal Growth 504 (2018) 17–25

A. Haque et al.

Fig. 4. (a) Plots of the cumulative thermal linear expansion coefficient for diamond, silicon, AlN and sapphire for comparison. The slope of each curve corresponds to the thermal expansion coefficient. (b) The thermal strain generated in the diamond thin film on different substrates (Si, AlN and Sapphire) at different temperature.

presence of biaxial stress, the Raman peak of polycrystalline diamond splits into singlet and doublet states and shifts from the equilibrium position. The linear expression between the stress and the peak shift due to the singlet and doublet states can be obtained by averaging the stress-peak shift relations. The shift of the singlet and doublet states can be expressed according to the following expressions developed by Ager and Drory [29].

Δνs [cm−1] = −0.93σ [GPa]

(2)

Δνd [cm−1] = −2.61σ [GPa]

(3)

At low-stress region, typically 0 – 8 GPa, the peak splitting is not resolved and a weighted shift relation, Δν = (1/3)νs + (2/3)νd, can be used to calculate the overall stress present in the diamond film. Therefore, for a diamond peak with no split, we can express the relation between the stress and the overall peak-shift as follows:

Δν [cm−1] = vs−v0 = −1.62σ [GPa] Fig. 5. Raman spectrum of the HFCVD diamond shows a sharp characteristic peak at 1333.85 cm−1. The inset shows the Raman spectrum of Q-carbon.

(4)

where v0 is the equilibrium peak position of the diamond in the unstressed state (1332.5 cm−1), vs corresponds to the position of the observed Raman peak in the stressed state. In this study, the deconvolution of the characteristic Raman peak into two individual peaks of singlet and doublet states reliably is very difficult due to comparatively small FWHM. The splitting of diamond peak to singlet and doublet only becomes observable at σ > 8–9 GPa [30]. From the position of the Raman peak, we obtained ∼1.14 GPa compressive stress in the diamond film. Such a low value of compressive stress is due to the presence of intermediate Q-carbon layer which nullifies the effect of thermal strain. Other researchers have observed a compressive stress of ∼3.7 GPa, even grown at low temperature (525 °C) in their studies [24]. Our stress value is the lowest ever obtained from diamond thin films grown on a single crystal sapphire substrate. The following expression can be used to determine the stress in the diamond film.

diamond at 1334.35 cm−1, which refers to the crystalline diamond growth. This peak occurs due to the zone center Raman active TO(X) optical phonon mode of diamond, which has triply degenerate T2g symmetry [25]. On the contrary, the synthetic diamond with defects present in the crystal is characterized by several additional features which have been attributed to different types of sp2 and sp3 defects. The natural diamond displays a sharp band at ∼1332.5 cm−1 during the Raman spectroscopy [26]. The linewidth of this peak is observed to be ∼2 cm−1. Usually, this characteristic peak of diamond shifts from its equilibrium position towards right due to the compressive stress in the film, as explained below. The peak position of the Raman signal is determined from the energy of the zone center optical phonons. The frequency of the first order phonon bands and, therefore, the relative position of the first order Raman line is related to the mass of the relevant atoms and the length and stiffness of the bonds in the crystal. Therefore, the position of the first order line will be sensitive to parameters that changes these parameters. The equilibrium position of the atoms is changed under a stressed condition. Stress creates a contraction of the lattice, and the interatomic forces are changed due to the subtle anharmonicity of the bonds. This change in interatomic forces alters the phonon vibrational frequencies, which can be characterized by the change of the Raman peak position. The shift in the peak position measured for a thin film can be used to estimate the stress using hydrostatic [27], uniaxial [28] or biaxial [29] models. Researchers determined relationships between the observed Raman peak shifts and stress values in the framework of biaxial stress model to study the stress in the diamond thin films on different substrates [30]. Under the

σth = dσ / dTX ΔT = [E /(1−ν )](αr −αs ) X ΔT = E (εd−εs )/(1−ν )

(5)

where E is the Young's modulus (1143 GPa) and ν is the Poisson ratio (0.07) of diamond, and εd and εs are temperature depended thermal expansions for the diamond film and the substrate, respectively. According to this expression, the maximum thermal stress that could be generated during the deposition of diamond under the chosen deposition conditions is obtained to be ∼ −5.1 GPa. According to Eq. (4), this value would result in ∼8.28 cm−1 blue shift in the Raman peak. However, in our study, the original peak position of the Raman spectrum shows a much less shift. It is believed that the use of the Q-carbon on the sapphire substrate plays a crucial role to reduce the effect of thermal stress. The Raman signal from a crystal of finite size also exhibits a smaller Raman shift and has a larger full width at half maximum (FWHM) than 21

Journal of Crystal Growth 504 (2018) 17–25

A. Haque et al.

positions and FWHM of the XRD peaks imply very good crystalline quality of the deposited diamond film with little strain. The XRD spectrum does not show any peak other than sapphire (0 0 0 6) peaks and diamond characteristic peaks, which eliminates the presence of any other phases in the sample. Stresses and strains play significant roles in determining the characteristics of the fundamental properties of thin film heterostructures, e.g., diamond on sapphire in this study. A generalized relationship between the stress and strain in a cubic (isotropic) crystal, a crystalline diamond thin film in this study, can be expressed by σzz = λεxx + λεyy + (λ + 2μ)εzz, where λ is the Lamé constant, and μ is the shear modulus [36]. In thin film heterostructures, there is no stress perpendicular to the film (free surface), so σzz = 0. Using the generalized expression one can obtain the following relationship between the strain in the perpendicular direction εzz with in-plane strains, εxx and εyy, and the Poisson’s ratio of the film (ν).

that from an infinite crystal lattice. In this study, the FWHM of the diamond peak is determined to be ∼8.5 cm−1, which is smaller compared to that of diamond grown by other techniques on sapphire. In different studies researchers have determined FWHMs of around 15 cm−1 and 12 cm−1 in the stressed diamond on sapphire grown by microwave enhanced chemical vapor deposition and low-temperature low-pressure electron–cyclotron–resonance CVD techniques and obtained good crystallinity in the diamond films [24,31]. Nevertheless, the FWHM is greater than the instrumental resolution of ∼1.2 cm−1. The broadening of this peak also refers to the generation of high compressive stresses in the diamond film [29]. Combined contributions of some other factors, such as microstructural defects in the film, small domain size of the diamond crystals, and sp2-bonded impurities, are also responsible for the broadening of this peak [32]. It is worthwhile to know that non-diamond carbon phases are codeposited with diamond during the growth of diamond thin films. These carbon- containing phases are also very sensitive to the Raman spectroscopy and can be characterized by their distinctive Raman signature, for example, graphite displays a single Raman band at 1580 cm−1. Amorphous carbon or diamond-like carbon is characterized by a very broad Raman band centered in-between 1450 cm−1 and 1550 cm−1. The exact shape and position of this band depend on the deposition process and the parameters used during the deposition [33]. In the as-deposited diamond film, the broad hump in the Raman spectra was centered at ∼1510 cm−1. Usually, the intensity of this peak in CVD diamonds on the sapphire substrate has been observed to be very high compared to the characteristic diamond peak [24,34]. The observed small intensity of amorphous carbon-related peak in this study refers to the excellent quality of the diamond thin film. Researchers also observed a small peak at ∼1150 cm−1 in CVD diamond on sapphire, which promotes the trace sp3-bonded carbon nanoclusters [13,33]. Nevertheless, the vanishing intensity of this peak in this study refers to the better quality of the diamond thin film grown by this process. The inset in Fig. 5 shows the Raman spectrum of Q-carbon film and gives the percentage of sp3 carbon as approximately 80–82%. The rest of the carbon remains in sp2 form.

εzz /(εxx + εyy ) = −ν /(1−ν )

(6)

Since εxx = εyy in the basal plane for the diamond thin film, we simplify Eq. (6) as εzz/εxx = −2ν/(1 − ν). We have measured the strain (εzz) around (1 1 1) peak from the θ-2θ X-ray diffraction scan and obtained an in-plane lattice strain of around 1.45 × 10−3. We can calculate the in-plane stress in the diamond film using this value of the planar lattice strain by using the following expression (Eq. (7)) [37].

σ = 2εxx μ (1 + ν )/(1−ν )

(7)

where ν is the Poisson’s ratio. The calculated stress in the diamond film is ∼1.57 GPa which is consistent with the stress value determined from Raman analysis. 3.5. Discussions There are significant advantages of using Q-carbon as a suitable intermediate layer for the fabrication of HFCVD diamond, which include: (i) the diamond tetrahedra present in the Q-carbon structure act as nucleation sites during the diamond growth, thus eliminating the requirement of a preexisting seed layer, (ii) no carbide formation is required at the film-substrate interface for adhesion, (iii) the melting point of Q-carbon is very high, which helps us to achieve the hightemperature CVD diamond growth and attain high growth rate (0.5 μm/ h) with improved quality of diamond thin films, (iv) there is no phase transition of the substrate or Q-carbon under the chosen deposition conditions, and (v) most importantly the Q-carbon helps to bypass the stringent requirement of low thermal misfit between the diamond and the transparent substrate. Although researchers [38,39] have proposed that the abstraction of hydrogen from the hydrocarbon mixture in the gaseous phase sublimated and mixed in the plasma at high-temperature is one of the possible mechanisms for diamond nucleation on the substrate with a preexisting seed layer, it is important to discuss the possible mechanism of diamond nucleation and growth on Q-carbon coated sapphire. It has been shown that Q-carbon consists of 80–85% sp3bonded carbon which are closely packed in the form of individual diamond tetrahedra with a packing efficiency of > 80% [16]. These diamond tetrahedra in Q-carbon act as the embryo for diamond nucleation. Therefore, the hydrocarbon species might decompose into nanosized diamond particles after releasing all of its hydrogen atoms at the Q-carbon surface. A schematic of the diamond tetrahedron is shown in Fig. 7(a) and the illustration of the diamond nucleation is shown in Fig. 7(b). The edge carbon atoms in the diamond tetrahedra or in a cluster of the diamond tetrahedra (also called Q-diamond) present in the Q-carbon are very active for the diamond nucleation due to the dangling bond characteristics, which act as active sites for the diamond nucleation. Subsequently, the growth of large-area diamond thin film takes place. In the previous study, it was shown that the Q-carbon is well adhered to the sapphire substrate due to the nonequilibrium liquid phase mediated growth from the molten carbon film which is quenched from a super undercooled state, which is created by using high-power

3.4. XRD We have conducted X-ray diffraction analysis to determine the crystallinity, phase composition, stress, and quality of diamond films grown on a sapphire substrate coated with Q-carbon. Fig. 6 represents a typical θ-2θ X-ray diffraction pattern obtained from the CVD diamond film on the c-sapphire substrate. This spectrum confirms the characteristic diamond peaks at (2θ) 43.94° and 75.42° corresponding to diamond {1 1 1} and {2 2 0} reflections, respectively [35]. The

Fig. 6. X-ray two theta scan of the HFCVD diamond thin film on Q-carbon coated c-sapphire. 22

Journal of Crystal Growth 504 (2018) 17–25

A. Haque et al.

Fig. 7. (a) An individual diamond tetrahedron. (b) The formation of new diamond tetrahedra (red) bonded with the preexisting diamond tetrahedra present in the Qcarbon during the nucleation process. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

formation of cracks, peeling off/delamination of the diamond film during cool down. Therefore, in this particular system (diamond on sapphire), the thermal mismatch plays a crucial role during film deposition as the thermal mismatch is comparable to the intrinsic part of the mismatch in the film-substrate system. However, in this present experiment, we overcame the effect of this huge thermal mismatch by integrating the Q-carbon layer on the sapphire substrate using a highly non-equilibrium pulsed laser annealing technique before the deposition of CVD diamond. The information about the maximum stress that can be accommodated without delamination of the film can serve as a platform to quantitatively assess the diamond adhesion to the substrate. The HFCVD diamond on Q-carbon/sapphire promotes the accommodation of thermal stress between the diamond film and the substrate and inhibits the delamination of the film from the substrate. Table 1 represents important substrate and film parameters (thermal expansion coefficient, Raman peak shift, measured stress, thermal stress) of different substrates which have been used for the growth of diamond thin film in different studies. These studies have shown that the diamond thin film can sustain a compressive stress of 2.835GPa without delamination [49]. Our measured compressive stress by Raman and XRD analyses are found to be considerably less than this value, which explains the large-area formation of diamond thin films on Q-carbon coated sapphire without delamination. The usefulness of diamond thin film on a substrate is considerably reduced if the diamond film has poor

nanosecond laser pulses [16]. The Q-carbon consists of ∼80% sp3 carbon and the rest is sp2 bonded carbon. The sp3 carbon atoms form numerous diamond tetrahedra in the dense Q-carbon structure, which act as the nucleation sites for the diamond growth [16]. As a result, the adhesion of Q-carbon with sapphire is enhanced, and its sublimation or evaporation at high-temperature is reduced. This increases the probability for a very stable Q-carbon, consisting of numerous diamond tetrahedra, to remain on the surface and act as seed/embryo for diamond nucleation. The residual amorphous carbon layer might get etched off due to the reaction with the highly active hydrogen radicals during the HFCVD process. Generally, deposition of CVD diamond on any substrate proceeds in two stages, namely, nucleation and growth. Over the years, different diamond nucleation methods have been developed. It is well established that diamond nucleation on any smooth non carbide forming surface such as mirror-polished sapphire substrate is very difficult. The nucleation density of diamond on smooth sapphire surface without any conditioning/pretreatment is only about 104 cm−2 which results in scattered isolated individual diamond crystallites on the substrate [13]. Diamond nucleation can be enhanced by scratching the substrate surface with diamond [40], SiC [41], cubic-BN [42], Cu or stainless steel [43], etc. Although substrate scratching can improve the diamond nucleation, it can cause serious damage to the periodic arrangement of the surface structure/atoms and thus degrades the quality of heterostructure. Coating the substrate surface with graphite [11], amorphous carbon [11,44], diamond-like carbon [45–47], C60, and mechanical oil [48] have also been employed to enhance the diamond nucleation. However, these type of coatings may result the formation of a graphitic interlayer in-between the diamond film and the substrate causing poor adhesion. This technique of diamond nucleation on Q-carbon can help to overcome the aforementioned obstacles. The investigation of the nucleation process reveals a narrow window parameters for highquality diamond nucleation. The crucial parameters are the substrate temperature, filament temperature, methane concentration, distance between the filaments and the substrate, and the nucleation/growth time. After nucleation the diamond crystals mainly increase laterally and after a sufficient amount of growth time the coalescence of the crystallites takes place. It is very important to obtain information about the thermal stress for different applications of diamond thin films such as optical protective coatings, fabrication of thin membranes, cutting tools and anti-friction coatings that undergo high mechanical loads frequently, and so on. A large difference in the thermal behavior of the diamond thin film and a substrate can lead to problems related to cracking and delamination. Besides the intrinsic stress, the difference in the thermal expansion can lead to additional stress during cool down. In a diamond-sapphire system, this strain is compressive due to the negative value of the percent of thermal misfit. The high thermal strain in the diamond film plane on sapphire during the growth results in the

Table 1 Comparison among different parameters (thermal expansion coefficient α, Raman peak shift Δν, measured stress σmeasured, thermal stress σth) of the diamond thin film grown on different substrates and by different techniques [13,24,30,31].

23

Substrate

α (10−6 K−1)

Δν, cm−1

σmeasured, GPa

σth, GPa

SiO2 Si SiC WC-Co (6%) Mo Sapphire (hightemperature CVD deposition) Sapphire (low temperature CVD deposition) Sapphire (Pt coated) Q-carbon/Sapphire (Present study)

0.5 2.5 2.8 5.2 5.3 7.5

+2.3 −1.0 −1.6 −3.8 −4.1 Film delaminated

+2.1 −0.25 −1.0 −2.0 −2.6 5.1

7.5

−1.0 to −3.7 1.2–1.6 0.9–2.6 3.8–6.2 6.5–6.7 0.7 & 16 (peak splitting) 5.5

3.2 (from Raman)

3.7

9.1 7.5

18 2.35

– 1.57 (from XRD) 1.14 (from Raman)

5.9 5.1

Journal of Crystal Growth 504 (2018) 17–25

A. Haque et al.

the conventional preexisting seed of nano/micro diamond on the substrate to deposit a continuous diamond thin film. The deposited CVD diamond films have a very small residual stress, compared to the other CVD diamond on sapphire, which has been determined by the Raman spectroscopy and XRD analysis. Less residual stress in the diamond film helps to deposit films with larger areas and higher thicknesses without any crack or delamination over the entire surface of the substrate. This process of growing CVD diamond on sapphire can open a new frontier towards the fabrication of large-area crystalline diamond film at hightemperature on other transparent substrates as well.

adhesion with the substrate. Our novel approach to fabricate large-area diamond films also improves the adhesion with the sapphire substrate. This has been achieved by the laser irradiation of the amorphous carbon layer to form a liquid state. The liquid carbon is quenched in a super undercooled state to form Q-carbon which is very adherent to the substrate. The diamond tetrahedra embedded in the Q-carbon are very adherent to the substrate due to the liquid phase mediated growth. Such adhesive diamond tetrahedra not only act as nucleation seeds for diamond growth but also in a way improve the overall adherence of the film with the substrate. To examine the extent of adhesion qualitatively, we have used the scotch tape method. No peeling of the diamond film was observed in the SEM images even after using the scotch-tape for 3 times. Among different applications of diamond on sapphire some notable examples are wear-resistant coatings, cutting tool applications, and windows in space applications. Additionally many microelectronic features require patterned polycrystalline diamond thin films [4]. Plasma etching, chemical etching, and/or selective growth are some of the patterning techniques used in the modern microelectronic technology. However, one of the major obstacles in obtaining patterned polycrystalline diamond is the highly resistive property/inertness of diamond to most of the chemical solutions. Therefore, diamond is very difficult to pattern by chemical etching and finding a suitable mask material, even if it could be chemically etched, to withstand the etchant is also very difficult. So far, reactive plasma etching of diamond with oxygen has resulted in limited success due to the poor etch selectivity between the diamond and the mask material [50]. Researchers also used alternative/non-conventional ions such as xenon ion and nitrogen dioxide to pattern diamond thin films [51]. However, these processes require complex equipment with low yield. Selective growth is another viable alternative process for patterning the diamond. Since diamond is grown over 800 C under extreme environment, we cannot use photoresist as a mask material. Therefore, researchers have used scratching or roughening techniques to grow diamond selectively. Such kind of selective regions on the substrates are obtained- (i) by using ethyl alcohol in an ultrasonic agitator along with standard photolithography and subsequent etching of the uncovered surface by Ar+ ion beam up to 500–600 Å in depth [52], (ii) by using pulsed laser irradiation on silicon and copper substrate utilizing a shadow mask or by scanning a focused laser beam, which may not work for sapphire [53], or (iii) by using diamond paste to damage the substrate, photolithographically patterning, and finally plasma etching the wafer in the undesired diamond growth areas to remove the damage or reducing the activation sites to restrain the growth of diamond [4]. However, these techniques result in damaging or scratching the substrate surface, where the damage includes large density of sharp edges, strains, pitts, microscratches, mechanical defects etc. Therefore, these techniques for selective diamond growth are not very useful for microelectronic device processing. The present technique of growing diamond could be an attractive alternative as Q-carbon can be selectively grown on single crystal sapphire substrate using the controlled laser irradiation process with a suitable optical setup [54]. This could be a desirable approach to selectively pattern diamond on sapphire and various other substrates with minimum stress in the film and without creating any damage on the substrate.

Acknowledgements This research was supported by the Grant No. NSF-DMR-1735695 and ARO-W911NF-17-1-0596. The authors would like to acknowledge the use of analytical instrument facility (AIF) at North Carolina State University, which is supported by the State of North Carolina and National Science Foundation. We also wish to thank the Fan Family Foundation Distinguished Chair Endowment for Professor J. Narayan. References [1] I. Friel, S.L. Geoghegan, D.J. Twitchen, G.A. Scarsbrook, Development of high quality single crystal diamond for novel laser applications, 2010, pp. 783819. https://doi.org/10.1117/12.864981. [2] B. Mesa, S. Magonov, Novel diamond/sapphire probes for scanning probe microscopy applications, J. Phys. Conf. Ser. 61 (2007) 770, https://doi.org/10.1088/ 1742-6596/61/1/154. [3] R.L. Fork, W.W. Walker, R.L. Laycock, J.J.A. Green, S.T. Cole, Integrated diamond sapphire laser, Opt. Express 11 (2003) 2532–2548, https://doi.org/10.1364/OE.11. 002532. [4] R. Ramesham, T. Roppel, Selective growth of polycrystalline diamond thin films on a variety of substrates using selective damaging by ultrasonic agitation, J. Mater. Res. 7 (1992) 1144–1151, https://doi.org/10.1557/JMR.1992.1144. [5] M.D. Drory, J.W. Hutchinson, Diamond coating of titanium alloys, Science 263 (1994) 1753–1755, https://doi.org/10.1126/science.263.5154.1753. [6] C.H. Prescott, W.B. Hincke, Equilibrium between aluminum carbide and nitrogen at high temperatures, J. Am. Chem. Soc. 50 (1928) 3228–3237, https://doi.org/10. 1021/ja01399a011. [7] P.W. May, C.A. Rego, C.G. Trevor, E.C. Williamson, M.N.R. Ashfold, K.N. Rosser, N.M. Everitt, Deposition of diamond films on sapphire: studies of interfacial properties and patterning techniques, Diam. Relat. Mater. 3 (1994) 1375–1380, https:// doi.org/10.1016/0925-9635(94)90153-8. [8] A.V. Sumant, V.P. Godbole, S.T. Kshirsagar, Studies on hot filament chemical vapor deposition of diamond onto aluminum oxide, Mater. Sci. Eng. B. 39 (1996) L5–L9, https://doi.org/10.1016/0921-5107(96)01579-6. [9] Y. Mo, Y. Xia, W. Wu, A nucleation mechanism for diamond film deposited on alumina substrates by microwave plasma CVD1Supported by the National Natural Science Foundation of China. 1, J. Cryst. Growth 191 (1998) 459–465, https://doi. org/10.1016/S0022-0248(98)00141-9. [10] X. Qiao, O. Fukunaga, N. Shinoda, K. Yui, Enhancement of diamond nucleation on alumina substrates by the activated species transport method, Diam. Relat. Mater. 5 (1996) 1096–1102, https://doi.org/10.1016/0925-9635(96)00526-2. [11] V.V.S.S. Srikanth, Review of advances in diamond thin film synthesis, Proc. Inst. Mech. Eng. Part C J. Mech. Eng. Sci. 226 (2012) 303–318, https://doi.org/10.1177/ 0954406211422788. [12] K.G. Saw, I. Andrienko, A. Cimmino, P. Spizzirri, S. Prawer, J. du Plessis, Growth of diamond on α-(0001) sapphire substrates, Diam. Relat. Mater. 12 (2003) 1663–1669, https://doi.org/10.1016/S0925-9635(03)00189-4. [13] O. Ternyak, R. Akhvlediani, A. Hoffman, Study on diamond films with ultra high nucleation density deposited onto alumina, sapphire and quartz, Diam. Relat. Mater. 14 (2005) 323–327, https://doi.org/10.1016/j.diamond.2004.10.033. [14] K.G. Saw, J. du Plessis, Diamond growth on faceted sapphire and the charged cluster model, J. Cryst. Growth 279 (2005) 349–356, https://doi.org/10.1016/j. jcrysgro.2005.02.072. [15] Y.-C. Chen, L. Chang, Chemical vapor deposition of diamond on an adamantanecoated sapphire substrate, RSC Adv. 4 (2014) 18945–18950, https://doi.org/10. 1039/C4RA01042F. [16] J. Narayan, A. Bhaumik, S. Gupta, A. Haque, R. Sachan, Progress in Q-carbon and related materials with extraordinary properties, Mater. Res. Lett. 6 (2018) 353–364, https://doi.org/10.1080/21663831.2018.1458753. [17] A. Haque, J. Narayan, Electron field emission from Q-carbon, Diam. Relat. Mater. 86 (2018) 71–78, https://doi.org/10.1016/j.diamond.2018.04.008. [18] J. Narayan, A. Bhaumik, Novel phase of carbon, ferromagnetism, and conversion into diamond, J. Appl. Phys. 118 (2015) 215303, https://doi.org/10.1063/1. 4936595. [19] D. Das, R.N. Singh, A review of nucleation, growth and low temperature synthesis of diamond thin films, Int. Mater. Rev. 52 (2007) 29–64, https://doi.org/10.1179/ 174328007X160245.

4. Summary We have demonstrated a simple and effective method for the direct deposition of large-area crystalline diamond film on c-sapphire using Qcarbon as an intermediate layer by HFCVD. The results from high resolution SEM, Raman spectroscopy, and XRD analysis show that the Qcarbon assists the diamond nucleation and a high-quality large-area continuous CVD diamond film on sapphire can be obtained. The Qcarbon helps to minimize the effect of thermal stress in the diamond film on single crystal sapphire and nullifies the stringent requirement of 24

Journal of Crystal Growth 504 (2018) 17–25

A. Haque et al.

[38] T. Nagano, N. Shibata, Diamond synthesis by microwave-plasma chemical vapor deposition using CH3Cl and CH2Cl2 as carbon source, Jpn. J. Appl. Phys. 32 (1993) 5067, https://doi.org/10.1143/JJAP.32.5067. [39] J. Jiang, Y. Tzeng, Mechanisms of suppressing secondary nucleation for low-power and low-temperature microwave plasma self-bias-enhanced growth of diamond films in argon diluted methane, AIP Adv. 1 (2011) 042117, https://doi.org/10. 1063/1.3656241. [40] Y. Mitsuda, Y. Kojima, T. Yoshida, K. Akashi, The growth of diamond in microwave plasma under low pressure, J. Mater. Sci. 22 (1987) 1557–1562, https://doi.org/10. 1007/BF01132374. [41] A. Sawabe, T. Inuzuka, Growth of diamond thin films by electron-assisted chemical vapour deposition and their characterization, Thin Solid Films 137 (1986) 89–99, https://doi.org/10.1016/0040-6090(86)90197-5. [42] N. Fujimori, T. Imai, A. Doi, Characterization of conducting diamond films, Vacuum. 36 (1986) 99–102, https://doi.org/10.1016/0042-207X(86)90279-4. [43] C-. Chang, D.L. Flamm, D.E. Ibbotson, J.A. Mucha, Diamond crystal growth by plasma chemical vapor deposition, J. Appl. Phys. 63 (1988) 1744–1748, https:// doi.org/10.1063/1.339912. [44] A.A. Morrish, P.E. Pehrsson, Effects of surface pretreatments on nucleation and growth of diamond films on a variety of substrates, Appl. Phys. Lett. 59 (1991) 417–419, https://doi.org/10.1063/1.105448. [45] J. Singh, M. Vellaikal, Nucleation of diamond during hot filament chemical vapor deposition, J. Appl. Phys. 73 (1993) 2831–2834, https://doi.org/10.1063/1. 353034. [46] K.V. Ravi, C.A. Koch, H.S. Hu, A. Joshi, The nucleation and morphology of diamond crystals and films synthesized by the combustion flame technique, J. Mater. Res. 5 (1990) 2356–2366, https://doi.org/10.1557/JMR.1990.2356. [47] K.V. Ravi, C.A. Koch, Nucleation enhancement of diamond synthesized by combustion flame techniques, Appl. Phys. Lett. 57 (1990) 348–350, https://doi.org/10. 1063/1.103688. [48] S. Yugo, T. Kanai, T. Kimura, T. Muto, Generation of diamond nuclei by electric field in plasma chemical vapor deposition, Appl. Phys. Lett. 58 (1991) 1036–1038, https://doi.org/10.1063/1.104415. [49] J.J. Gracio, Q.H. Fan, J.C. Madaleno, Diamond growth by chemical vapour deposition, J. Phys. Appl. Phys. 43 (2010) 374017, https://doi.org/10.1088/00223727/43/37/374017. [50] T.J. Whetten, A.A. Armstead, T.A. Grzybowski, A.L. Ruoff, Etching of diamond with argon and oxygen ion beams, J. Vac. Sci. Technol. A. 2 (1984) 477–480, https:// doi.org/10.1116/1.572598. [51] N.N. Efremow, M.W. Geis, D.C. Flanders, G.A. Lincoln, N.P. Economou, Ion-beamassisted etching of diamond, J. Vac. Sci. Technol. B Microelectron. Process. Phenom. 3 (1985) 416–418, https://doi.org/10.1116/1.583276. [52] K. Hirabayashi, Y. Taniguchi, O. Takamatsu, T. Ikeda, K. Ikoma, N. IwasakiKurihara, Selective deposition of diamond crystals by chemical vapor deposition using a tungsten-filament method, Appl. Phys. Lett. 53 (1988) 1815–1817, https:// doi.org/10.1063/1.99789. [53] J. Narayan, X. Chen, Laser patterning of diamond films, J. Appl. Phys. 71 (1992) 3795–3801, https://doi.org/10.1063/1.350892. [54] A. Haque, J. Narayan, Stability of electron field emission in Q-carbon, MRS Commun. (2018) 1–9, https://doi.org/10.1557/mrc.2018.172.

[20] H. Sumiya, N. Toda, S. Satoh, Growth rate of high-quality large diamond crystals, J. Cryst. Growth 237–239 (2002) 1281–1285, https://doi.org/10.1016/S00220248(01)02145-5. [21] W.R.L. Lambrecht, C.H. Lee, B. Segall, J.C. Angus, Z. Li, M. Sunkara, Diamond nucleation by hydrogenation of the edges of graphitic precursors, Nature 364 (1993) 607–610, https://doi.org/10.1038/364607a0. [22] W.M. Yim, R.J. Paff, Thermal expansion of AlN, sapphire, and silicon, J. Appl. Phys. 45 (1974) 1456–1457, https://doi.org/10.1063/1.1663432. [23] S.B. Qadri, C. Kim, E.F. Skelton, T. Hahn, J.E. Butler, Thermal expansion of chemical vapor deposition grown diamond films, Thin Solid Films 236 (1993) 103–105, https://doi.org/10.1016/0040-6090(93)90651-5. [24] R.K. Singh, D.R. Gilbert, J. Laveigne, Growth of adherent diamond films on optically transparent sapphire substrates, Appl. Phys. Lett. 69 (1996) 2181–2183, https://doi.org/10.1063/1.117158. [25] J. Filik, J.N. Harvey, N.L. Allan, P.W. May, J.E.P. Dahl, S. Liu, R.M.K. Carlson, Raman spectroscopy of nanocrystalline diamond: an ab initio approach, Phys. Rev. B. 74 (2006) 035423, https://doi.org/10.1103/PhysRevB.74.035423. [26] S.A. Solin, A.K. Ramdas, Raman spectrum of diamond, Phys. Rev. B. 1 (1970) 1687–1698, https://doi.org/10.1103/PhysRevB.1.1687. [27] D.S. Knight, W.B. White, Characterization of diamond films by Raman spectroscopy, J. Mater. Res. 4 (1989) 385–393, https://doi.org/10.1557/JMR.1989.0385. [28] M. Yoshikawa, G. Katagiri, H. Ishida, A. Ishitani, M. Ono, K. Matsumura, Characterization of crystalline quality of diamond films by Raman spectroscopy, Appl. Phys. Lett. 55 (1989) 2608–2610, https://doi.org/10.1063/1.101951. [29] J.W. Ager, M.D. Drory, Quantitative measurement of residual biaxial stress by Raman spectroscopy in diamond grown on a Ti alloy by chemical vapor deposition, Phys. Rev. B. 48 (1993) 2601–2607, https://doi.org/10.1103/PhysRevB.48.2601. [30] V.G. Ralchenko, E.D. Obraztsova, K.G. Korotushenko, A.A. Smolin, S.M. Pimenov, V.G. Pereverzev, Stress in thin diamond films on various materials measured by microraman spectroscopy, MRS Online Proc. Libr. Arch. 383 (1995), https://doi. org/10.1557/PROC-383-153. [31] T. Tachibana, Y. Yokota, K. Kobashi, M. Yoshimoto, Heteroepitaxial growth of {111}-oriented diamond films on platinum{111}/sapphire{0001} substrates, J. Cryst. Growth 205 (1999) 163–168, https://doi.org/10.1016/S0022-0248(99) 00223-7. [32] L. Bergman, R.J. Nemanich, Raman and photoluminescence analysis of stress state and impurity distribution in diamond thin films, J. Appl. Phys. 78 (1995) 6709–6719, https://doi.org/10.1063/1.360495. [33] R.J. Nemanich, J.T. Glass, G. Lucovsky, R.E. Shroder, Raman scattering characterization of carbon bonding in diamond and diamondlike thin films, J. Vac. Sci. Technol. A. 6 (1988) 1783–1787, https://doi.org/10.1116/1.575297. [34] A. Samoto, S. Ito, A. Hotta, T. Hasebe, Y. Ando, A. Sawabe, T. Suzuki, Investigation of heterostructure between diamond and iridium on sapphire, Diam. Relat. Mater. 17 (2008) 1039–1044, https://doi.org/10.1016/j.diamond.2008.02.007. [35] M. Kamo, Y. Sato, S. Matsumoto, N. Setaka, Diamond synthesis from gas phase in microwave plasma, J. Cryst. Growth. 62 (1983) 642–644, https://doi.org/10.1016/ 0022-0248(83)90411-6. [36] J.P. Hirth, J. Lothe, Theory of Dislocations, McGraw-Hill, New York, London, 1968. [37] J. Narayan, Recent progress in thin film epitaxy across the misfit scale (2011 Acta Gold Medal Paper), Acta Mater. 61 (2013) 2703–2724, https://doi.org/10.1016/j. actamat.2012.09.070.

25